I'm in the very early stages of designing a kick bin section for a 3way mid to based around the B&C 15ndl76. I'm aiming to use it for between 100-300hz. I'm hoping for an average sensitivity of 102db -3db in the pass band. It will be going below a sonitus mr8xt and Beyma cd10fe (one cabinet).

I know it's a gargantuan task but, I've seen it done on this forum a few times before (Davey's Mid Top for example) . If anyone has any pointers I'd be very happy to hear them, I'm not easily offended. 

I'm very comfortable with the woodworking aspects and expecting to be making a fair few prototypes, as well as spend a good deal of time on Solidworks. 

Here's some of the hornresp sims so far in 4pi:

Any advice would be greatly appreciated! 

Folded horn, but only two folds. Straight horn like the coffin bin would be impractical. 

Ideally the cabinet would be trapezoidal. Height around the 1200 mark, width  552 narrowing to 344 for a 28 degree splay angle with 445mm depth.

Hopefully it won't be too honky. 

Hi luthier. I think to model s1 correctly you got to take the whole baffle into account not just the driver aperture. So to accomodate a 15" driver i would think it would need to be atleast 40x40cm, so probably closer to 1600cm2?

Also i dont know if it would reach 300hz with a fold, from what ive read bandbass horns like that dont like to play that high. A straight horn like hornresp the model would probably be better .

I''ve designed my own straight horn for this driver based on the coffin bin project posted by another member. However, with no folds it ended up being over 800mm in length.

There are a fair few BPH in production that play 300hz and above. I'm hoping to be able to the same.

Something I have found useful in the past when testing designs like this is to compare the impedance plots between the simulation and the real cabinet.  Best way to analyse how close to the theory you've got.

I'm quite interested to follow this because firstly I'm quite familiar with those midrange horns but also I've tried a few similar low frequency designs in the past and not come up with anything worthy of building into a complete cabinet.  You are definitely right to expect a few rounds of prototyping!

Yea your right in saying the es18 is modeled with s1 of 800cm2 but thats because s1 in that simulation is taken from where the front chamber ends (s2 in your simulation). It might be worth trying to sim using the fromt chber value in hornresp and removing the first section of the horn to compare sims.

There you go! Ive wondered myself a few times, good to know the two meathods agree.

need to reach 400hz not 300hz. I've been able to do this with a relatively short horn but I'm loosing all my low end between 100hz and 140hz. I've tried adding a port to compensate but I can't seem to control the xmax and loose significant sensitivity. 

If I make my S1 smaller I get a better result but anything under 300 seems to get hornresp angry. I imagine it's down to the compression ratio. I wonder If anyone uses a phase plug to correct this in a bph or just a very small throat.

Let's start with simulating a single segment horn and incorporate directivity, without that you're only seeing half the story really.

What makes you say you've got an excursion problem, because I'm pretty sure that you shouldn't have one.

Best regards Johan

Here's a look at the directivity, I was a bit shocked about the spl increase between 200hz and 400hz, but I couldn't see any lobes in the directivity pattern.

Here's the sim of the MR8xt:

So I've now increased the horn length pretty much doubling it, but it is starting to produce the results I wanted to see. 

Here's a few pictures of the process.

Could anyone shed some light on the S1 section? I'm planing on getting the 125mm segment length chasing down the back of the cabinet as I really don't want to make it any deeper! 

If I was to fill the chamber I would need to fill this amount! 

 Here's a few pictures of the different iterations of the cabinet, last one being most recent.

As a plus it also seems to control the excursion. 

Thanks for the tips! I'm now able to get the same response on axis with 20mm of lining at an FR of 4000, I do have a 4-5db peak off axis between 10-30 degrees between 200hz and 500hz. 

Not sure If this needs to be resolved or not?

Excuse the F1 style front panel, it wasn't intentional, but it makes it really easy to get the 15" in and out! 

This is the cabinet matching the last and most recent plot, I've got the excursion under control down to 101hz and the 15 just about makes it to 370hz. So not absolutely perfect for meeting the sonitus mr8xt but acceptable. 

In the current format I can always go ported later in real life testing and make various chambers to play with tuning. 

I will also experiment with some different drivers at some point in Hornresp. 

I have simulated the full horn in Hornresp as it is drawn in CAD and haven't been able to notice and resonances at this time. If you have any suggestions on how I might go about simulating this differently or more accurately? in hornresp or any other software I'd be interested to know!?

My hope is that mouth area is large enough and cabinet construction rigid enough that resonances do not occur in the passband. I would fully expect to see them higher up but I only plan on using this section of the cabinet between 120 - 400hz. 

In an ideal world I would be able to get the horn with -3db to 80hz, at the movement I can get down to 100hz -3db but I run out of xmax at 105hz so plan on passing at 120hz to be safe. 

One question I have been racking my brain over recently is if this section of the vtc chamber is counted as the LRC or the S1 path length? At the moment (shown in grey) it is modelled as the lrc. The red line shows if I where to sim it as the s1.

Here are some simulations with different drivers:

Most simmed like the Faital pro hp1020 but the ndl88 and Oberton 15b450 looked slightly better than the ndl76. I'm trying to stick to neodymium drivers to keep the size and weight of the enclosure down. The tops will probably weigh 40-50kg plus so any weight saving is a plus. 

So far I have simmed: FP 15fh520, FP 15fhp1020, Beyma 15lw30, B&C NDL88, B&C NDL76, emi kappalite 3015lf, Oberton 15b450

B&C 15ndl76 - Grey 

Faital Pro hp1020 - black

Pd.153er - Black

Pd.153er - Grey

Pd.1550 - black

As it is with my current horn path length my pass band is limited to 112hz - 400hz within +/-3db which is 12hz higher than my ideal high pass and 100hz higher than my orginal low pass.

With the Oberton 15b450 and pd.153er I can gain slightly higher lowpass at around 500hz although this is not nessecary for my design requirements. However, given the low price of the Oberton, £121.45 (bluearan) I will definetley keep it as a possible future option. The NDl88 also simmed well.

The cabinet in it's current layout seems to like drivers with a small Vas, Bl around 20, large xmax and relativley large VC.

To gain a better response below 105hz without going into the xmax I need to increase the horn length, which I'm reluctant to do as it does not seem to provide any significant improvement. For a few extra hz - around 10 - I reduce the low pass and lower overall sensitivity.

Porting gets me down to 85hz but the trade-off with sensitvity and lpf doesn't seem worth it. My average sensitivity is around 100.5db and the xmax is still too high. 

In terms of concerns, the off-axis response is around 3db higher between 200-400hz the the standard plot and the response is a bit peaky at max spl, looks a little like the golden arches!  

My cabinet dimensions are  H x 1036mm W x  537mm D x 472mm which is very close to my orignal size parameters, which I'm very happy about. 

I also set out to acheive 102db sensitivy +/-3db in the pass band which I have managed. 

Here's my proposed crossover points, simmed in hornresp.

Hopefully this covers it. 

Hopefully this will help to reduce the forces on the cone, it also allows me to play down to 80hz, realistically I will only ever cross at 100hz + though.

In hornresp I've had a go adding filters and EQ, just wondering on what people think of this and if it is a step in the wrong direction. I know off-axis I have a 3db boost between 200-300hz.

Is it worth just protoyping and mucking around with filters and eq after measurements?

I'm just trying to get the design as good as I can before I prototype. It's doesn't cost me money doing this!  

this is 4pi.

Not to worry, I've been trying to get that down to 9L for a while now, currently it's at 11.8L

Intrested to know peoples thoughts on the ports being in the horn mouth?

Sim is on axis so down about 3db on the high end, sims much better than any previous sim from my models. No real displacment issues down to 80hz and goes up to about 330hz with out too much lost spl but I'd like to try and push it higher. I'm thinking of filling sections of the chamber with styrofoam during the testing.

Wish I had your drawing skills in Inventor, plan for today is to put some time aside and hopefully not jut look at it with a glazed look as usual, been pulling my hair out trying to get to grips with it.

With regards to ports in the mouth, I don't think you will see any real effect as so close to the mouth though David did include the ability to model it in HR some time ago, I tried to model a box that had the port at the throat, that didn't show any useful results, very different to the measured response, since he added that feature I have managed to lose all my measured data of frequency response and  impedance plots, it could well be that I have just misused HR though - more likely than not.

Wish I had your drawing skills in Inventor

Never actually spent any time on Inventor, looks like a very nice bit of software. I mostly use Solidworks, Rhino, Autocad and Grasshopper.


plan for today is to put some time aside and hopefully not jut look at it with a glazed look as usual, been pulling my hair out trying to get to grips with it.

With regards to ports in the mouth, I don't think you will see any real effect as so close to the mouth though David did include the ability to model it in HR some time ago, I tried to model a box that had the port at the throat, that didn't show any useful results, very different to the measured response, since he added that feature I have managed to lose all my measured data of frequency response and impedance plots, it could well be that I have just misused HR though - more likely than not.

Thanks for the help info on the port I was hoping that was the case.

Here's what happens when the volume of the VRC is reduced (grey is 15L red is (9.4L)

If I increase the size to it's current size with no volume reduction in any way it is 18.8L when the 9.281L of driver displacement has been removed. Maybe if I stick a big ferrite driver in there I can loose some more volume! 

(grey 15L, Red 18.8L)

First picture is the volume of the shape with nothing subtracted/intersected.

Second is after intersection.

Looks like you were right! Think this means I need to start check my late night CAD modeling when I'm a litttle more rested!  

Also apart from feeling a bit stupid I'm quite happy, I was only 9ml ! over what the manufacturers claimed displacment!

Thank you for spotting my error! 

Hoping to be able to conisder all of the blue section as my VRC and not yet part of my horn path.

I'll probably stick it in another seperate thread on speakerplans.

(not too bad for just over 125L, I don't think)

I'm increasing the throat size to reduce the compression ratio, I'm also moving away from the porting. My thoughts being if I can't get the results I want with a sealed chamber there is nothing stoping me adding them at a later date. 

Removing the ports has a few noticable implications it increases output above 130hz and causes it to drop drastically below 120hz. To increase output in the lower region I have to reduce the chamber volume to 40L. 

Interestingly if I model the horn as a single section (without ports) the off axis response is relativley smooth and very sensitive. 

 

https://www.diyaudio.com/forums/multi-way/279476-3d-modeling-tips-tricks.html" rel="nofollow - https://www.diyaudio.com/forums/multi-way/279476-3d-modeling-tips-tricks.html

RoadRunnersDust wrote:  Interesting design/build to follow! Have you tried sim'ing the Beyma 15LW30 in it?

So before the big update on the project here's a screen grab from solidworks; you can see a couple of fouling issues but this is pretty much the final layout for prototyping. 

I've got a Boxford system too. The software sucks donkey balls! I feel your pain!

Not sure about releasing plans, it would make a very difficult DIY build due to the compound angles inherent with a bph in a trapezoidal cab.